Dave

1K posts

Dave

Dave

@LLdavedave

🇮🇪➡️🇺🇸➡️🇦🇺

Brisbane شامل ہوئے Mart 2009
2.6K فالونگ650 فالوورز
Dave
Dave@LLdavedave·
@eoinmcc Coding related tasks or something else?
English
1
0
0
24
Eoin McCarthy
Eoin McCarthy@eoinmcc·
I've been looking for a way to manage larger chunks of work I want to offload to Claude. I might be a bit late to the game but Paperclip looks pretty interesting - github.com/paperclipai/pa…
English
3
0
2
70
Dave
Dave@LLdavedave·
@nicbstme I found the messaging interface is a strange unexpected unlock, before using it I said you can use the Claude / ChatGPT app on iOS how is this different but it is
English
0
0
0
422
Nicolas Bustamante
Nicolas Bustamante@nicbstme·
Why do I need openclaw? Am I missing out? I run Claude Code from my phone with /remote-control. It has access to my filesystem, professional and personal Drive and emails with GogCLI, my sms etc. It seems like 80% of the value prop without the security headache and token cost.
English
166
7
438
113.9K
Dave
Dave@LLdavedave·
@anothercohen I was waiting for this. You never disappoint
English
0
0
1
2.8K
Alex Cohen
Alex Cohen@anothercohen·
I was fired from Block today. I was the PM in charge of changing the default tip option on the Square terminal to start at 40%. Jack replaced me with an AI agent that decides which tip amount to show based on your age, weight, and race. If anyone is hiring product managers, please let me know!
English
213
212
5.6K
997.4K
Dave
Dave@LLdavedave·
Hey @nikitabier big fan of your work. Will operation kill the bots tackle this? They’re getting more frequent
Dave tweet media
English
0
0
0
23
Dave
Dave@LLdavedave·
I hadn’t heard of @Nanoclawsol before this, love the idea of skills instead of config and the claw rewriting itself based on your desired setup. super cool!
Andrej Karpathy@karpathy

Bought a new Mac mini to properly tinker with claws over the weekend. The apple store person told me they are selling like hotcakes and everyone is confused :) I'm definitely a bit sus'd to run OpenClaw specifically - giving my private data/keys to 400K lines of vibe coded monster that is being actively attacked at scale is not very appealing at all. Already seeing reports of exposed instances, RCE vulnerabilities, supply chain poisoning, malicious or compromised skills in the registry, it feels like a complete wild west and a security nightmare. But I do love the concept and I think that just like LLM agents were a new layer on top of LLMs, Claws are now a new layer on top of LLM agents, taking the orchestration, scheduling, context, tool calls and a kind of persistence to a next level. Looking around, and given that the high level idea is clear, there are a lot of smaller Claws starting to pop out. For example, on a quick skim NanoClaw looks really interesting in that the core engine is ~4000 lines of code (fits into both my head and that of AI agents, so it feels manageable, auditable, flexible, etc.) and runs everything in containers by default. I also love their approach to configurability - it's not done via config files it's done via skills! For example, /add-telegram instructs your AI agent how to modify the actual code to integrate Telegram. I haven't come across this yet and it slightly blew my mind earlier today as a new, AI-enabled approach to preventing config mess and if-then-else monsters. Basically - the implied new meta is to write the most maximally forkable repo and then have skills that fork it into any desired more exotic configuration. Very cool. Anyway there are many others - e.g. nanobot, zeroclaw, ironclaw, picoclaw (lol @ prefixes). There are also cloud-hosted alternatives but tbh I don't love these because it feels much harder to tinker with. In particular, local setup allows easy connection to home automation gadgets on the local network. And I don't know, there is something aesthetically pleasing about there being a physical device 'possessed' by a little ghost of a personal digital house elf. Not 100% sure what my setup ends up looking like just yet but Claws are an awesome, exciting new layer of the AI stack.

English
0
0
0
77
Dave
Dave@LLdavedave·
@steipete @Austen ha love that you found time to see this and reply
English
0
0
1
171
Dave
Dave@LLdavedave·
I never got into crypto, even when so many friends were all in. Over the last few weeks whenever I talk to people IRL I usually end up talking about AI. Then later I wonder if this is how all the crypto people felt when ICOs and NFTs were all the rage. Definitely believe something is happening
John Palmer@johnpalmer

x.com/i/article/2021…

English
0
0
0
99
Dave
Dave@LLdavedave·
I never got into crypto, even when so many friends were all in. Over the last few weeks whenever I talk to people IRL I usually end up talking about AI. Then later I wonder if this is how all the crypto people felt when ICOs and NFTs were all the rage. Definitely believe something big is happening
Matt Shumer@mattshumer_

x.com/i/article/2021…

English
0
0
1
87
Dave
Dave@LLdavedave·
@jackfriks @mikecallaghan I get the 'better for me'. In my case I think I'm more productive with Cursor than others because I have got comfortable with the Cursor UI (despite all their changes!). I've tried Claude Code, Codex and @opencode and I keep coming back to Cursor.
English
1
0
2
118
jack friks
jack friks@jackfriks·
i can’t believe there are people still using cursor who have never tried claude code…
English
239
14
751
80.4K
Dave
Dave@LLdavedave·
@EXM7777 Are you using skills.sh? If not is there somewhere else you're finding useful skills or are you just building your own?
English
0
0
0
114
Machina
Machina@EXM7777·
claude skills are the most underrated investment you can make right now... they're the natural evolution to prompts, and soon everyone will be looking to collect them what makes this so good: > easy to understand for complete beginners > easy to build because claude does it for you > the model recognizes when a skill is needed based on context that last part is the game changer for me... you're not manually loading prompts anymore: the model sees what you're doing and injects the right context, SOPs, frameworks... automatically output quality AND speed go up without extra work i love that more providers are calling them skills too (pretty sure anthropic named them first) makes it clear this is becoming a universal setup and in a world where new models drop every 3 weeks and beat everyone? being able to transfer skills from one platform to another is really convenient go build some skills... it might just change your life
English
45
22
526
33.6K
Dave
Dave@LLdavedave·
Earlier this month I started using some skills from @steipete 's github.com/steipete/agent… which have really helped my workflow. On the hunt for more I discovered skills.sh but I'm quickly realising the number of installs a skill has is not necessarily an indicator of quality. Is there anywhere to find vetted skills?
English
0
0
0
33
Dave
Dave@LLdavedave·
@kr0der @steipete revisiting this as I'm looking into skills again for a new project. Do you use skills.sh , if not do you have any other recommendations (beyond searching github) for finding skills?
English
0
0
0
11
Anthony
Anthony@kr0der·
i basically copied almost all of the AGENTS md file and edited the personal info to my own. i didn't include stuff i don't use such as the tailscale part, private ops, sparkle keys, etc also copied skills from the skills folder based on what i actually use, into ~/.codex/skills. e.g. i copied oracle, nano-banana-pro, etc the scripts folder i copied into ~/Projects/agent-scripts, same as the docs folder (except slash-commands, i copied those into ~/.codex/prompts
English
2
0
1
681
@levelsio
@levelsio@levelsio·
Got this cute new air quality sensor in Brazil because we forgot ours at home It was $50, kinda cheap so we were already sus Then I tried it at home and in mall and it showed same exact values (hmmmm 🤔) Then I searched fake air sensor on YouTube and found my exact one 😂
@levelsio tweet media@levelsio tweet media@levelsio tweet media
English
145
18
1.3K
179.5K
Dave
Dave@LLdavedave·
this is wild
TBPN@tbpn

Clawdbot creator @steipete describes his mind-blown moment: it responded to a voice memo, even though he hadn't set it up for audio or voice. "I sent it a voice message. But there was no support for voice messages. After 10 seconds, [Moltbot] replied as if nothing happened." "I'm like 'How the F did you do that?'" "It replied, 'You sent me a message, but there was only a link to a file with no file ending. So I looked at the file header, I found out it was Opus, and I used FFmpeg on your Mac to convert it to a .wav. Then I wanted to use Whisper, but you didn't have it installed. I looked around and found the OpenAI key in your environment, so I sent it via curl to OpenAI, got the translation back, and then I responded.'"

English
0
0
1
65
Dave
Dave@LLdavedave·
@levelsio Tried the plastic bag test and it immediately jumped to 3k. Doesn't mean it is accurate but it is at least detecting extremes
Dave tweet media
English
1
0
0
145
Dave
Dave@LLdavedave·
@levelsio I have the exact same one in white bought on Amazon AU. Now you have my suspicious of the readings. Will try the bag with bad air test as suggested below.
Dave tweet media
English
2
0
2
863
Dave
Dave@LLdavedave·
@ArmanHezarkhani I know some non tech people not on X who’ve successfully built products and internal tools using it. I’ve helped them out a few times when they’ve been blocked/stuck. Too abstract and closed garden for me but I can see why non technical folks enjoy it
English
0
0
2
87
Arman Hezarkhani
Arman Hezarkhani@ArmanHezarkhani·
Has anyone used Base44? I never hear about it on tech twitter, but I ALWAYS get ads for it Whats going on?
English
5
0
3
982